[Ifeffit] Running FEFF 8 in Artemis

[Matt Newville]

Hi Jose,

On Sat, Sep 14, 2019 at 1:53 PM José C. Conesa <jcconesa at icp.csic.es> wrote:

> Dear Matt,
>
> I exchanged with you on this matter some messages in September 2017, and
> after that I exchanged also other messages with Bruce Ravel in December
> 2017 and January 2018. He told me that he was very busy with its new
> beamline and had no time to adapt Artemis so that it could use feff 8.5L.
> This was of course long ago. If Artemis can now use the output of feff
> 8.5L, I'm glad to know and will try it. I presume that it will be enough to
> tell Artemis the location of this feff code.
>
Artemis can use the *outputs* of Feff8: the feffnnnn.dat files are the same
format as from Feff6.   I think that what you were having trouble with was
having Artemis run Feff8 for you.  I do not believe that has changed,
though I'm not entirely sure why that was not working.  But you can run
Feff8L outside of Artemis and then import the resulting path files.

> On the other hand, I do not use normally Larch, although I know about it.
> Are there advantages in using Larch instead of Artemis?
>
Larch comes with Feff8l, which is sort of what you were asking about. Larch
does not (yet?) have a GUI for running Feff or for fitting a sum of Feff
paths to a set of EXAFS data.  It does have Python functions to do all
these steps so that you could write a simple script to perform a fit.
Artemis can use Larch as its underlying EXAFS fitting engine.  Artemis does
not (yet?) do everything that Larch can do (fit in wavelet space, for
example), but Larch does not (yet?) provide "fuzzy paths", and you would
have to work those out in your own script.

Hope that helps,

--Matt
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://millenia.cars.aps.anl.gov/pipermail/ifeffit/attachments/20190916/965e565c/attachment.html>